SignificanceLabel-free, two-photon excited fluorescence (TPEF) imaging captures morphological and functional metabolic tissue changes and enables enhanced understanding of numerous diseases. However, noise and other artifacts present in these images severely complicate the extraction of biologically useful information.AimWe aim to employ deep neural architectures in the synthesis of a multiscale denoising algorithm optimized for restoring metrics of metabolic activity from low-signal-to-noise ratio (SNR), TPEF images.ApproachTPEF images of reduced nicotinamide adenine dinucleotide (phosphate) (NAD(P)H) and flavoproteins (FAD) from freshly excised human cervical tissues are used to assess the impact of various denoising models, preprocessing methods, and data on metrics of image quality and the recovery of six metrics of metabolic function from the images relative to ground truth images.ResultsOptimized recovery of the redox ratio and mitochondrial organization is achieved using a novel algorithm based on deep denoising in the wavelet transform domain. This algorithm also leads to significant improvements in peak-SNR (PSNR) and structural similarity index measure (SSIM) for all images. Interestingly, other models yield even higher PSNR and SSIM improvements, but they are not optimal for recovery of metabolic function metrics.ConclusionsDenoising algorithms can recover diagnostically useful information from low SNR label-free TPEF images and will be useful for the clinical translation of such imaging.
The potential to differentiate between diseased and healthy tissue has been demonstrated through the extraction of morphological and functional metrics from label-free, two-photon images. Acquiring such images as fast as possible without compromising their diagnostic and functional content is critical for clinical translation of two-photon imaging. Computational restoration methods have demonstrated impressive recovery of image quality and important biological information. However, access to large clinical datasets has hampered advancement of denoising algorithms. Here, we seek to demonstrate the application of denoising algorithms on depth-resolved two-photon excited fluorescence (TPEF) images with specific focus on recovery of functional metabolic metrics. Datasets were generated through the collection of images of reduced nicotinamide adenine dinucleotide (phosphate) (NAD(P)H) and flavoproteins from freshly excised rat cheek epithelium. Image datasets were patched across depth, generating 1012, 256-by-256 patches. A well-known U-net architecture was trained on 6628 low-signal-to-noise-ratio (SNR) patches from a previously collected large dataset and later retrained on a smaller 620 low-SNR patches dataset before being validated and evaluated on 88 and 304 low-SNR patches, respectively, using a structural similarity index measure (SSIM) loss function. We demonstrate models trained on larger datasets of human cervical tissue could be used to successfully restore metabolic metrics with an improvement in image quality when applied to rat cheek epithelium images. These results motivate further exploration of weight transfer for denoising of small clinical two-photon microscopy datasets.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.